Sample

Buse/61

Human rating uses the original dataset scale.

DatasetBuse
Human3.78
Lines6

Scores

MethodScoreScaleDataset rank
(0–100)
Posnett0.3590–1; higher is more readable
57.6Δ 15.2
Dorn0.8420–1; higher is more readable
93.9Δ 21.2
Mi0.6540–1; higher is more readable
67.7Δ 5.1
LLOC3Logical lines; fewer is more readable
88.9Δ 16.2
DeepSeek V4 Pro11.670–20; mean of three runs
23.2Δ 49.5
GPT-6.1 Sol16.670–20; mean of three runs
53.5Δ 19.2
BTTF0.4530–1; higher is more readable
38.4Δ 34.3
Human3.781–5; original rating
72.7

Views

Download official ZIP ↗

Select the downloaded ZIP or its extracted folder. Files stay in your browser.

Loading feature diagnosis…

Source regions

Feature contributions

Feature contributions

FeatureValueContribution
(log-odds)
Byte entropyHow varied the bytes in the source text are. Repeated bytes give a lower value; a more even mixture gives a higher value.4.821−7.232
Halstead volumeHow much symbolic information the code contains. It combines the total number of operators and values with the number of distinct ones.140−4.62
Source linesThe number of non-empty source lines.6+2.4

Intercept 8.87 + feature contributions = -0.582 (log-odds).
Score = sigmoid(-0.582) = 0.359.

Feature contributions

FeatureValueContribution
(log-odds)
Identifier layout (horizontal)How rapidly the visual pattern of identifier characters changes from left to right. It measures spatial frequency, not the meaning of the names.44.857+1.765
Comment layout frequencyHow frequently the layout switches between lines containing only comments or whitespace and lines containing code.7+0.788
Indentation frequencyHow rapidly indentation changes from line to line.1−0.703
Number layout frequencyHow rapidly the number of numeric values per line changes through the code.1+0.398
Identifier layout (vertical)How rapidly the visual pattern of identifier characters changes from top to bottom.5.565−0.332
Operator visual areaThe share of the code’s visual area occupied by operators such as +, =, and *.0.121−0.232
Comment layout (horizontal)How rapidly the visual pattern of comment characters changes from left to right.45−0.078

Intercept 0.07 + feature contributions = 1.677 (log-odds).
Score = sigmoid(1.677) = 0.842.

Score0.654
Score3
Run 1

The code is fairly readable with clear intent: mapping a return type to an opcode and emitting it. Structure is logical and nesting is minimal, limited to a flat else-if chain. Action granularity is reasonable, with one main operation per line. However, readability is slightly reduced by the inline comment '//long' which is ambiguous and could be clearer. The dangling else-if structure also assumes prior context, making the snippet harder to evaluate in isolation. Overall, the code is concise and natural to follow, but minor clarity issues prevent a higher score.

Run 2

The code snippet is relatively short and each action is generally on its own line, which helps granularity. However, readability is hurt by inconsistent indentation (starting with an extra tab for the 'else if' chain), vague variable names ('cv', 'opcode'), and magic strings ('F', 'J') that require domain knowledge. The comment '//long' offers a minor explanation but the intent could be clearer. The structure is straightforward but the formatting reduces the natural reading flow.

Run 3

The code snippet is short and functional but suffers from readability issues. The indentation is inconsistent and off from the left margin, suggesting it's a fragment of a larger conditional block. The different `equals` comparisons each check a different type against a letter like 'F' or 'J', and although there's a comment clarifying 'J' means long, the purpose of the mapping isn't immediately clear. The else-if chain is flat and reasonably simple, but because of the lack of surrounding context, the logic feels ambiguous. Action granularity is fine—one meaningful operation per line—but the final `cv.visitInsn(opcode);` call is separated from its dependency, which weakens the reading flow. Overall, the code is concise but not especially clear or self-explanatory.

Run 1

Code Structure: The conditional branches logically map return-type codes to opcodes, followed by instruction emission, though the preceding branches are missing. Nesting: Flat and easy to follow, with no deep nesting. Clarity of intent: Opcode names and visitInsn suggest return-instruction generation, but "F" and "J" require familiarity with type descriptors; only "J" is explained. Code length: Concise, with no unnecessary statements. Action granularity: Each assignment and method call occupies its own line. Reading flow: The branches naturally lead to the emission call, although heavy indentation, extra spacing, and omitted braces slightly reduce readability.

Run 2

Code Structure: The else-if branches logically map return types to opcodes, followed by instruction emission, though the preceding branches are missing. Nesting: Flat conditional branches are easy to follow. Clarity of intent: Opcode names clearly indicate return instructions, but the single-letter type descriptors require domain knowledge, and only the long case has an explanatory comment. Code length: Concise with no unnecessary statements. Action granularity: Each assignment and method call occupies its own line. Reading flow: The progression from opcode selection to emission is natural, although excessive indentation, inconsistent spacing, and omitted braces slightly reduce readability.

Run 3

Code Structure: The else-if chain logically maps return-type descriptors to return opcodes, followed by instruction emission. Nesting: The snippet is shallow and easy to follow. Clarity of intent: Opcode names and visitInsn convey the purpose, though the descriptor strings "F" and "J" require domain knowledge; only "J" is explained. Code length: Concise, with no unnecessary statements. Action granularity: Each assignment and method call occupies its own line. Reading flow: Natural top-to-bottom progression, slightly weakened by inconsistent spacing, omitted braces, and the snippet beginning mid-chain, which hides opcode initialization and other cases.